Kalman Filter-Based Chip Differential Blind Adaptive Multiuser Detection for Variably Mobile Asynchronous Underwater Multiuser Communications

(1)

University of Wollongong

Research Online

Faculty of Engineering and Information

Sciences - Papers: Part B

Faculty of Engineering and Information

Sciences

2018

Kalman Filter-Based Chip Differential Blind Adaptive Multiuser Detection for

Variably Mobile Asynchronous Underwater Multiuser Communications

Guang Yang

Qingdao University

Qinghua Guo

University of Wollongong, [email protected]

Defeng (David) Huang

University of Western Australia, [email protected]

Jingwei Yin

Harbin Engineering University

Maochun Zheng

Harbin Engineering University

Follow this and additional works at: https://ro.uow.edu.au/eispapers1

Part of the Engineering Commons, and the Science and Technology Studies Commons

Recommended Citation

Yang, Guang; Guo, Qinghua; Huang, Defeng (David); Yin, Jingwei; and Zheng, Maochun, "Kalman Filter-Based Chip Differential Blind Adaptive Multiuser Detection for Variably Mobile Asynchronous Underwater Multiuser Communications" (2018). Faculty of Engineering and Information Sciences - Papers: Part B. 1805.

https://ro.uow.edu.au/eispapers1/1805

Research Online is the open access institutional repository for the University of Wollongong. For further information contact the UOW Library: [email protected]

(2)

Kalman Filter-Based Chip Differential Blind Adaptive Multiuser Detection for

Variably Mobile Asynchronous Underwater Multiuser Communications

Abstract

To mitigate the inter-symbol interference (ISI), multiple access interference (MAI) and the effect of different Doppler shifts within a data package incurred in variably mobile asynchronous underwater multiuser communications (VMAUMC), a Kalman filter-based chip differential blind adaptive multiuser detection (D-KF-BAMUD) algorithm is proposed in this paper. The Kalman filter-based blind adaptive multiuser detection (KF-BAMUD) algorithm has been adopted to combat ISI and MAI in the scenario of static users, and underwater and under-ice multiuser communication experiments have been carried out to verify the effectiveness of the KF-BAMUD algorithm. To tackle the more challenging VMAUMC, a differential modulation and demodulation technique operating at the chip level is proposed to reduce the impact of Doppler shift residues after the Doppler compensation with an average Doppler frequency offset within each data package. The technique is combined with KF-BAMUD algorithm, leading to D-KF-BAMUD algorithm, which is able to effectively handle ISI, MAI and the effect of Doppler shift residues simultaneously in VMAUMC. This paper reports our experiments on VMAUMC carried out in Songhua River in April 2016 (7 users, data rate 7.87bit/s, up to 2m/s variably relative velocity in each data package, and horizontal distance 200m shallow-water channel between transducer and hydrophone), and the encouraging experimental results demonstrate the effectiveness of the proposed system and the D-KF-BAMUD algorithm.

Disciplines

Engineering | Science and Technology Studies

Publication Details

G. Yang, Q. Guo, D. Huang, J. Yin & M. Zheng, "Kalman Filter-Based Chip Differential Blind Adaptive Multiuser Detection for Variably Mobile Asynchronous Underwater Multiuser Communications," IEEE Access, vol. 6, (1) pp. 49646-49653, 2018.

(3)

Received July 13, 2018, accepted August 29, 2018, date of publication September 4, 2018, date of current version September 28, 2018.

Digital Object Identifier 10.1109/ACCESS.2018.2868475

Kalman Filter-Based Chip Differential Blind

Adaptive Multiuser Detection for Variably

Mobile Asynchronous Underwater

Multiuser Communications

GUANG YANG 1,2_{, QINGHUA GUO} 3_{, (Member, IEEE),}

DEFENG HUANG4_{, (Senior Member, IEEE),}

JINGWEI YIN2_{, AND MAOCHUN ZHENG}2

1_{School of Information and Control Engineering, Qingdao University of Technology, Qingdao 266033, China}

2_{Acoustic Science and Technology Laboratory, College of Underwater Acoustic Engineering, Harbin Engineering University, Harbin 150001, China} 3_{School of Electrical, Computer and Telecommunications Engineering, University of Wollongong, Wollongong, NSW 2522, Australia}

4_{School of Electrical, Electronic and Computer Engineering, The University of Western Australia, Perth, WA 6009, Australia} Corresponding author: Jingwei Yin ([email protected])

This work was supported in part by the National Natural Science Foundation of China under Grants 61771271 and 61631008, in part by the Heilongjiang Provincial Natural Science Foundation for Outstanding Youth, China, under Grant JC2017017, in part by the Fok Ying-Tong Education Foundation, China, under Grant 151007, in part by the Shandong Provincial Natural Science Foundation, China, under Grants ZR2016FP06, ZR2017MF024, and ZR2018MF002, in part by the Project of Shandong Province Higher Educational Science and Technology Program under Grant J18KA315, and in part by the Opening Fund of Acoustics Science and Technology Laboratory, China. The work of D. Huang was supported by the Australian Research Council Discovery Project under Grant DP110100736.

ABSTRACT To mitigate the inter-symbol interference (ISI), multiple access interference (MAI), and the

effect of different Doppler shifts within a data package incurred in variably mobile asynchronous underwater multiuser communications (VMAUMC), a Kalman filter-based chip differential blind adaptive multiuser detection (D-KF-BAMUD) algorithm is proposed in this paper. The Kalman filter-based blind adaptive multiuser detection (KF-BAMUD) algorithm has been adopted to combat ISI and MAI in the scenario of static users, and underwater and under-ice multiuser communication experiments have been carried out to verify the effectiveness of the KF-BAMUD algorithm. To tackle the more challenging VMAUMC, a differential modulation and demodulation technique operating at the chip level is proposed to reduce the impact of Doppler shift residues after the Doppler compensation with an average Doppler frequency offset within each data package. The technique is combined with the KF-BAMUD algorithm, leading to the D-KF-BAMUD algorithm, which is able to effectively handle ISI, MAI, and the effect of Doppler shift residues simultaneously in VMAUMC. This paper reports our experiments on VMAUMC carried out in Songhua River in April 2016 (seven users, data rate 7.87 bit/s, up to 2 m/s variably relative velocity in each data package, and horizontal distance 200-m shallow-water channel between transducer and hydrophone), and the encouraging experimental results demonstrate the effectiveness of the proposed system and the D-KF-BAMUD algorithm.

INDEX TERMS Inter-symbol interference, multiple access interference, Doppler shift residues, variably

mobile underwater multiuser communications.

I. INTRODUCTION

Underwater acoustic channels are hostile, which are charac-terized by low carrier frequency, narrow bandwidth, severe multipath delay, and large Doppler shift [1]. Research and experiments in the literature have been mainly focused on static underwater multiuser communications, i.e., both the

transducers and the hydrophones of the users are assumed to be static [2]–[17]. A typical application example is shown in Fig.1 (a), where a fixed underwater network communicates to a static autonomous underwater vehicle (AUV). In this scenario, two sensors transmit data to the AUV simultane-ously, and there is no Doppler shift because both sensors and

49646

2169-3536 2018 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

(4)

G. Yang et al.: D-KF-BAMUD for VMAUMC

FIGURE 1. (Color online) Comparison between application examples of this paper and some references

(a) Communication scenario in [2]–[10] (b) Communication scenario in [11]–[14] (c) Communication scenario 1 in this paper (d) Communication scenario 2 in this paper.

AUV are static. Another important scenario of underwater multiuser communications is variably mobile asynchronous underwater multiuser communications (VMAUMC) where the relative movement between transducers and hydrophones has to be considered. The communication environment of VMAUMC is more hostile due to the variable Doppler shifts within each data package incurred by the variably relative velocity of the transducer and the hydrophone.

To the best of our knowledge, only three organizations have carried out research and experiments on mobile underwa-ter multiuser communications with approximately constant velocities (the relative movement between the transducer and the hydrophone has constant velocities), which include Mas-sachusetts Institute of Technology (MIT), Scripps Institution of Oceanography, and University of Connecticut. An appli-cation example is shown in Fig.1 (b), where three AUVs communicate with a fixed sensor simultaneously. In this case, one AUV is static, the second is moving away from the sensor at a constant velocity and the third moves towards the sensor at another constant velocity. In 2013, Stojanovic in MIT

carried out research on underwater 2-user mobile communi-cation experiments in Kauai and the coast of Massachusetts in USA. In the first experiment, the relative velocity between the transducer and hydrophone of the expected user is 1.5m/s with a horizontal distance of 2km. In the second experiment, the relative velocity between the transducer and hydrophone of the expected user is 1m/s with a horizontal distance ranging from 0.5km to 4.5km [18]. In 2013, Zhou at the University of Connecticut conducted research on underwater 2-user mobile experiments in the coast of Massachusetts in USA. The rel-ative velocities between the transducer and hydrophone of the expected user are 1.2m/s and 1m/s with a horizontal distance between 0.5km and 4.5km, respectively [19], [20]. In the same year, Song in Scripps Institution of Oceanography carried out research on underwater 2-user mobile communi-cation experiments in Kauai in USA. The relative velocity between transducer and hydrophone of the expected user is 1.3m/s with a horizontal distance of 1.3km [21]. We can see that the maximum relative velocity of all the experi-ments mentioned above is 1.5m/s. In addition, the orthogonal

(5)

frequency division multiplexing (OFDM) is adopted for the communication systems in the experiments. It is noted that, in the above works, the channel state information is acquired by using a separate channel estimator at the hydrophone side. However, when the transducer and (or) hydrophone of the expected user move with variable velocities, e.g., in VMAUMC, it is very difficult to obtain an accurate channel estimate, which may thereby result in performance degrada-tion. Hence non-coherent detection without channel estima-tion may be desirable.

This paper is focused on VMAUMC. Two application scenarios of VMAUMC are shown in Fig.1 (c), where three variably mobile AUVs communicate with a variably mobile AUV simultaneously, and Fig.1 (d), where two variably mobile AUVs and a fixed sensor communicate to a floated buoy simultaneously. Fig.1 (c) shows three variably mobile transducers and one variably mobile hydrophone, where their velocities change freely. In Fig.1 (d), two variably mobile AUVs and a fixed sensor are transmitters, and the receiver is a floated buoy, where the AUVs’ velocities change freely. We have designed a Kalman filter-based blind adaptive mul-tiuser detection (KF-BAMUD) algorithm to deal with the inter-symbol interference (ISI) and multiple access interfer-ence (MAI) incurred in asynchronous underwater multiuser communications [14]. In 2017, the KF-BAMUD algorithm was successfully employed for under-ice multiuser communi-cations with rising and diving interfering users [15]. We made attempts to apply the KF-BAMUD algorithm to VMAUMC and conducted experiments in Bohai Sea in 2015 and 2016 and compared the effectiveness of the KF-BAMUD algorithm for underwater and under-ice communications. Our experiments showed that the KF-BAMUD algorithm does not work well when the transducer and the hydrophone of the expected user are variably relatively mobile due to the variable Doppler shifts within each data package, which is induced by the relative variable movement between the trans-ducer and hydrophone of the expected user.

In this work, we propose to use a differential modulation and demodulation technique at the chip level for the under-water communication system to mitigate the impact of the different Doppler shift residues in each data package after the Doppler compensation with an average Doppler frequency offset. For VMAUMC, although Doppler shifts of two adja-cent chips are not exactly the same, but they are very close because the time duration of the adjacent chips is very short. Therefore, Doppler shifts of two adjacent chips can be consid-ered approximately the same when using differential modula-tion and demodulamodula-tion at the chip level, which facilitates the use of the KF-BAMUD algorithm for VMAUMC and leads to a Kalman filter-based chip differential blind adaptive mul-tiuser detection (D-KF-BAMUD) algorithm. In April 2016, we carried out VMAUMC experiments with a horizontal dis-tance of 200m shallow-water channel between the transducer and hydrophone in the Songhua River, where the number of users is 7, the data rate is 7.87bit/s, and the variably relative velocity is up to 2m/s. The encouraging experimental results

are reported in this paper, which verify the effectiveness of the proposed system and the D-KF-BAMUD algorithm.

The paper is organized as follows. The system and algo-rithm designs are presented in Section II. Field experiments are reported in Section III. Finally, conclusions are drawn in Section IV.

II. MULTIUSER DETECTION FOR VMAUMC

The chip-level differential modulation and detection based spread spectrum underwater multiuser communication sys-tem for VMAUMC is shown in Fig.2, where the syssys-tem includes L users, and each user is assigned a unique Gold sequence for spreading. Without loss of generality, we take User1 as the expected user.

A. TRANSMITTERS WITH CHIP-LEVEL DIFFERENTIAL MODULATION

The transmitters are shown in Fig.2 (a), The information bits to be transmitted by User1 are denoted by {b1(n) , n ≥ 1},

which take +1 or 1 with equal probability independently. The spreading sequence for User1 is denoted by s1 =

[s1(1), s1(2), . . . , s1(Ns)]T, where Ns is the spreading gain.

After spreading, the chip sequence corresponding to the n-th information bit of User1 can be represented as

b1n= b1(n) s1, n ≥ 1 (1)

where b1n=[b1n(1), . . . , b1n(Ns)]T. Then, the chip

differen-tial operation is performed, yielding the differendifferen-tially coded bit

d1n(i) = b1n(i) ⊕ d1n(i − 1), 2 ≤ i ≤ Ns (2)

where ⊕ denotes the xor operation and d1n(1) = b1n(1). The

signal is then modulated with a carrier, i.e., the transmitted signal can be represented as

S1n(i) = d1n(i)ejωct, 1 ≤ i ≤ Ns (3)

whereωc=2πfcand fcis carrier frequency.

B. MULTIUSER DETECTION AT RECEIVER

It is known that underwater acoustic communication suf-fers from multipath propagation, especially in shallow water, which induces severe ISI. We adopt a simple strategy, where the signals from distinct paths of a user are treated as signals transmitted by some virtual users. That is, the receiver treats the ISI and MAI in the same way. We use K to represent the total number of virtual users.

Let the Doppler shift for virtual Userk be1fk, i.e.,1ωk =

2π1fk, and the corresponding path gain be Ak, then the

received signal in underwater variably mobile multiuser com-munications can be represented as [21]

r(i) = A1d1n(i)ej(ωc+1ω1)t+ K X k=2 Akdkn(i)ej(ωc+1ωk)t+ v(i) (4) 49648 VOLUME 6, 2018

(6)

FIGURE 2. (Color online) Chip-level differential modulation and detection multiuser communication system for VMAUMC (a) Transmitters in a multiuser communication system (b) Receiver in a multiuser communication system.

where v(i) denotes the white Gaussian noise, and we delib-erately separate the signal from User1 from the sum as we assume User1 is the expected user (the same way can be used to treat other users). As shown in Fig.2 (b), the differential demodulation is performed firstly, and the received chips are delayed by one chip duration Tc and then multiplied with

the original chip sequence, i.e. (5), where I0 = A1d1n(i −

1)ej(ωc+1ω1)(t−Tc)_I_{(i) + A}

1d1n(i)ej(ωc+1ω1)tI(i−1)+I (i−1)I (i)

and I (i) =

K

P

k=2

Akdkn(i)ej(ωc+1ωk)t + v(i). Taking the real

part of (5), as shown at the bottom of this page, we have (6), as shown at the bottom of this page. Then the signal is low pass filtered, which yields (7), as shown at the bottom of this page, where I00represents the interference plus noise after low pass filtering. As the Doppler shifts of two adjacent chips are approximately the same,1ω1and cos [(ωc+1ω1)Tc] can be

r(i)r(i − 1) = A1d1n(i)ej(ωc+1ω1)tA1d1n(i − 1)ej(ωc+1ω1)(t−Tc)

+ A1d1n(i − 1)ej(ωc+1ω1)(t−Tc)I(i)A1d1n(i)ej(ωc+1ω1)tI(i − 1) + I (i − 1)I (i)

= A1d1n(i)ej(ωc+1ω1)tA1d1n(i − 1)ej(ωc+1ω1)(t−Tc)+ I0 (5) Re {r(i) r (i − 1) } = A2₁d1n(i)d1n(i − 1) cos {(ωc+1ω1)t} cos {(ωc+1ω1) (t − Tc)} + I0

= A2₁[b1n(i) ⊕ d1n(i − 1)] d1n(i − 1) ·1 2{cos [(ωc+1ω1)Tc] + cos [2(ωc+1ω1)t − (ωc+1ω1)Tc]} + I 0 (6) r0(i) = A2₁[b1n(i) ⊕ d1n(i − 1)] d1n(i − 1) 1 2cos [(ωc+1ω1)Tc] + I 00 (7) VOLUME 6, 2018 49649

(7)

FIGURE 3. (Color online) Flow chart of the D-KF-BAMUD algorithm.

regarded as constants. Let kd =cos [(ωc+1ω1)Tc], then we

have r0(i) = A2₁[b1n(i) ⊕ d1n(i − 1)d1n(i − 1)] kd 2 + I 00 = A2₁[b1n(i) ⊕ 1] kd 2 + I 00 = −kdA 2 1 2 b1n(i) + I 00 (8) Define y0(n) = [r0(1), r0(2), . . . , r0(Ns)]T and I00 = [I00(1), I00(2), . . . , I00(Ns)] T

. We may use the spreading sequence s1to extract the transmitted bits of User1 from (8)

by the following correlation y0_(n)_{, s} 1 = sT1 −kdA2₁ 2 b1n+s T 1I 00 =sT₁−kdA 2 1 2 b1(n) s1+s T 1I 00 (9) However, it is noted that, the interference sT

1I

00 _{may be}

very significant because the orthogonality between sT₁ and I00 cannot be guaranteed, and the performance can be degraded with the increase of the number of users and their amplitudes. Hence, we propose to apply the blind adaptive multiuser detection technique [22]. The detector can be represented as

ˆ b1(n) = sgn c1(n), y0(n) = sgn cT₁(n)y0(n) (10) with c1(n) = s1−C1,nullx1(n) (11)

where, the columns of matrix C1,null span the null space

of s1, i.e. s1, C1,null = 0, and x1(n) is the adaptive

part of c1(n). This is inspired by the generalized side-lobe

canceller to minimize the impact of the interference. It is worth mentioning that, as s1is known, C1,nullcan be obtained

through off-line computation. The problem here is how to update c1(n) or x1(n), adaptively, which can be achieved

using the Kalman filtering [22] by building a state equation

and measurement equation. The VMAUMC system with chip level differential modulation can be treated as a slow time-varying system, and the optimal weight vector copt1(n +

1) ≈ copt1(n). So, the state equation for the optimal adaptive

weight vector xopt1(n) can be approximately written as:

xopt1(n + 1) = xopt1(n) (12)

Define

e(n) =c1, y0(n) = cT1(n)y 0

(n) (13) Substituting (11) into (13), we get:

e(n) = sT₁y0(n) − y0T(n)C1,nullx1(n) (14)

Define z(n) = sT₁y0_{(n) and H(n) = y}0T_(n)C

1,null. When x1(n)

is equal to xopt1(n), we get the measurement equation z(n) = H(n)xopt1(n) + eopt(n) (15)

Let variance of e(n) be ξ (n), and ξmin=E

n

e2_opt(n)o. From (14), the mean of eopt(n) is 0, and eopt(n) is a white

Gaussian noise with mean value 0 and varianceξmin. Now,

the multiuser detection for the VMAUMC system is trans-formed to the Kalman filtering problem with (12) as the state equation and (15) as the measurement equation. ˆxopt1(n)

can be found with the standard Kalman algorithm in (16), as shown at the bottom of this page, where we initialize ˆ

xopt1(0) = 0, P(1/0) = I and H (1) = y0T(1)C1,null.

The flow chart of the D-KF-BAMUD algorithm is shown in Fig.3. The iterative calculation is realized by the Kalman filter algorithm in (16).

III. 7-USER UNDERWATER VARIABLY MOBILE ASYNCHRONOUS COMMUNICATION EXPERIMENTS IN SONGHUA RIVER IN APRIL 2016

Due to the limited experimental resources, only one trans-ducer and one hydrophone were used for the field mea-surement. Similar to [18]–[21], the received signal for the multiuser communications is generated by superimposing the

H(n) = y0T(n)C1,null

K (n) = P(n/n − 1)HT(n)hH(n)P(n/n − 1)HT(n) +ξmin

i−1

P(n + 1/n) = [I − K(n)H(n)] P(n/n − 1)

ˆ

xopt1(n) = ˆxopt1(n − 1) + K (n)z(n) − H(n)ˆxopt1(n − 1)

c1(n) = s1−C1,nullxˆopt1(n) (16)

(8)

FIGURE 4. (Color online) Deployment, testing scene, channels and variably relative velocities of the VMAUMC

experiments in Songhua River (a) Field deployment (b) Testing scene (c) Channel impulse responses (d) Variably relative velocities of the estimated transmitted symbols.

collected data from the field measurement asynchronously. In each experiment, for the carrier demodulation of each user, we used an average phase to compensate the phase deviation and the average phase was estimated using a pilot sequence. The arrival time of the expected User1 is used as the benchmark, and the power of interfering users is artificially changed by scaling their signals so that different signal to interference ratios (SIRs) of User1 can be obtained. BPSK (Binary Phase Shift Keying) was used for all the experiments and the spreading gain 127. The SIR is defined as

SIR(dB) = 10log₁₀( PExpectedUser1

PInterferenceUsers

) (17) where PExpectedUser1 and PInterferenceUsers are the power of

the expected User1 and the total power of the other users, respectively. A 7-user experiment was carried out in Songhua River in April 2016. Deployment and testing scene of the VMAUMC experiments in Songhua River are shown in Fig.4 (a) and (b). The transducer was deployed at a fixed location, and it transmitted 7 packages of data to simulate 7 users, while the hydrophone moved away from the trans-ducer with about 3m/s. The original distance between the transducer and hydrophone was about 200m and the river depth was about 6m. Carrier frequency was 2.5kHz, sampling frequency was 16kHz and the bandwidth was 1kHz. Each user transmitted 60bit within a duration of 7.62s, i.e., the data

rate was 7.87bit/s. Channels of the experiments are shown in Fig.4 (c), where the blue stars represent the channel of the expected User1 and pink lines represent the channels of the interfering users. Variably relative velocities of the estimated symbols during one data packages (60 bit) are shown in Fig.4 (d), where the maximum velocity difference is about up to 2m/s.

The experimental results of the Songhua River multiuser communications are shown in Fig.5, where, besides D-KF-BAMUD, we also implement and test the performance of RLS-BAMUD and LMS-BAMUD with differential modu-lation (denoted by D-RLS-BAMUD and D-LMS-BAMUD). For the single user and multiuser cases, BERs are 0, and their constellation plots are shown in Fig.5 (a) and (b), where it can be seen that the performance of the D-KF-BAMUD algorithm is the best among the three algorithms for mitigating ISI and MAI. The output of the detectors with SIR = 2dB is shown in Fig.5 (c) and the BERs of the expected User1 versus SIR are shown in Fig.5 (d). We fixed SIR at 2dB and added random Gaussian white noise artificially to adjust the SNR of the expected User1 and tested 60 data packages. The BERs of the expected User1 versus SNR are shown in Fig.5 (e). It can be seen that the D-KF-BAMUD algorithm works well for VMAUMC, and it delivers the best performance for mit-igating ISI, MAI and Doppler shift residues among the three kinds of algorithms.

(9)

FIGURE 5. (Color online) VMAUMC experimental results in Songhua River (7 users) (a) Constellation plot of single user communication (b) Constellation plot of multiuser communication with SIR = 2dB (c) Detector output with SIR = 2dB (d) BERs of the expected User1 versus SIR (e) BERs of the expected User1 versus SNR.

IV. CONCLUSION

In this paper, we have proposed an underwater mul-tiuser communication system with chip level differential modulation and the corresponding D-KF-BAMUD algorithm for VMAUMC. In the proposed receiver, ISI is treated as MAI, so that ISI and MAI are handled simultaneously. The differential modulation at the chip level is used to reduce

the effect of Doppler shift residues within the data package after the Doppler compensation, which facilitates the use of a Kalman-based blind adaptive multiuser detection without channel estimation. The VMAUMC experiments have been carried out in Songhua River and the good experimental results demonstrate the effectiveness of the proposed system and algorithm.

(10)

REFERENCES

[1] P. Qarabaqi and M. Stojanovic, ‘‘Statistical characterization and computa-tionally efficient modeling of a class of underwater acoustic communica-tion channels,’’ IEEE J. Ocean. Eng., vol. 38, no. 4, pp. 701–717, Oct. 2013. [2] C. C. Tsimenidis, O. R. Hinton, A. E. Adams, and B. S. Sharif, ‘‘Under-water acoustic receiver employing direct-sequence spread spectrum and spatial diversity combining for shallow-water multiaccess networking,’’

IEEE J. Ocean. Eng., vol. 26, no. 4, pp. 594–603, Oct. 2001.

[3] M. Stojanovic and L. Freitag, ‘‘Multichannel detection for wideband underwater acoustic CDMA communications,’’ IEEE J. Ocean. Eng., vol. 31, no. 3, pp. 685–695, Jul. 2006.

[4] E. Calvo and M. Stojanovic, ‘‘Efficient channel-estimation-based mul-tiuser detection for underwater CDMA systems,’’ IEEE J. Ocean. Eng., vol. 33, no. 4, pp. 502–512, Oct. 2008.

[5] T. C. Yang and W.-B. Yang, ‘‘Interference suppression for code-division multiple-access communications in an underwater acoustic channel,’’

J. Acoust. Soc. Amer., vol. 126, no. 1, pp. 220–228, Jul. 2009.

[6] S. A. Aliesawi, C. C. Tsimenidis, B. S. Sharif, and M. Johnston, ‘‘Iterative multiuser detection for underwater acoustic channels,’’ IEEE J. Ocean.

Eng., vol. 36, no. 4, pp. 728–744, Oct. 2011.

[7] H. C. Song, W. S. Hodgkiss, W. A. Kuperman, T. Akal, and M. Stevenson, ‘‘Multiuser communications using passive time reversal,’’ IEEE J. Ocean.

Eng., vol. 32, no. 4, pp. 915–926, Oct. 2007.

[8] H. C. Song, J. S. Kim, W. S. Hodgkiss, W. Kuperman, and M. Stevenson, ‘‘High-rate multiuser communications in shallow water,’’ J. Acoust. Soc.

Amer., vol. 128, no. 5, pp. 2920–2925, Nov. 2010.

[9] S. E. Cho, H. C. Song, and W. S. Hodgkiss, ‘‘Successive interference cancellation for underwater acoustic communications,’’ IEEE J. Ocean.

Eng., vol. 36, no. 4, pp. 490–501, Oct. 2011.

[10] S. E. Cho, H. C. Song, and W. S. Hodgkiss, ‘‘Multiuser interference cancellation in time-varying channels,’’ J. Acoust. Soc. Amer., vol. 131, no. 2, pp. EL163–EL169, 2012.

[11] S. E. Cho, H. C. Song, and W. S. Hodgkiss, ‘‘Asynchronous multiuser underwater acoustic communications (L),’’ J. Acoust. Soc. Amer., vol. 132, no. 1, pp. 5–8, Jul. 2012.

[12] T. Shimura, H. Ochi, and H. Song, ‘‘Experimental demonstration of mul-tiuser communication in deep water using time reversal,’’ J. Acoust. Soc.

Amer., vol. 134, no. 4, pp. 3223–3229, 2013.

[13] T. C. Yang, ‘‘Spatially multiplexed CDMA multiuser underwater acoustic communications,’’ IEEE J. Ocean. Eng., vol. 41, no. 1, pp. 217–231, Jan. 2016.

[14] G. Yang, J. Yin, D. Huang, L. Jin, and H. Zhou, ‘‘A Kalman filter-based blind adaptive multi-user detection algorithm for underwater acoustic net-works,’’ IEEE Sensors J., vol. 16, no. 11, pp. 4023–4033, Jun. 2016. [15] J. Yin, G. Yang, D. Huang, L. Jin, and Q. Guo, ‘‘Blind adaptive multi-user

detection for under-ice acoustic communications with mobile interfering users,’’ J. Acoust. Soc. Amer., vol. 141, no. 1, pp. EL70–EL75, 2017. [16] L. Ma, S. Zhou, G. Qiao, S. Liu, and F. Zhou, ‘‘Superposition coding

for downlink underwater acoustic OFDM,’’ IEEE J. Ocean. Eng., vol. 42, no. 1, pp. 175–187, Jan. 2017.

[17] C.-U. Baek, J.-W. Jung, and D.-W. Do, ‘‘Study on the structure of an effi-cient receiver for covert underwater communication using direct sequence spread spectrum,’’ Appl. Sci., vol. 8, no. 1, p. 58, 2018.

[18] K. Tu, T. M. Duman, M. Stojanovic, and J. G. Proakis, ‘‘Multiple-resampling receiver design for OFDM over Doppler-distorted underwater acoustic channels,’’ IEEE J. Ocean. Eng., vol. 38, no. 2, pp. 333–346, Apr. 2013.

[19] Z. Wang, S. Zhou, J. Catipovic, and P. Willett, ‘‘Asynchronous multiuser reception for OFDM in underwater acoustic communications,’’ IEEE

Trans. Wireless Commun., vol. 12, no. 3, pp. 1050–1061, Mar. 2013. [20] J. Huang, S. Zhou, and Z. Wang, ‘‘Performance results of two iterative

receivers for distributed MIMO OFDM with large Doppler deviations,’’

IEEE J. Ocean. Eng., vol. 38, no. 2, pp. 347–357, Apr. 2013.

[21] S. E. Cho, H. C. Song, and W. S. Hodgkiss, ‘‘Multiuser acoustic com-munications with mobile users,’’ J. Acoust. Soc. Amer., vol. 133, no. 2, pp. 880–890, Feb. 2013.

[22] X.-D. Zhang and W. Wei, ‘‘Blind adaptive multiuser detection based on Kalman filtering,’’ IEEE Trans. Signal Process., vol. 50, no. 1, pp. 87–95, Jan. 2002.

GUANG YANG was born in 1981. He received the Ph.D. degree from Harbin Engineering University in 2016. From 2014 to 2016, he was a Visiting Scholar with The University of Western Australia. He is currently an Assistant Professor with the School of Information and Control Engineering, Qingdao University of Technology, China. His interests focus on underwater acoustic communi-cation and underwater acoustic physics.

QINGHUA GUO (S’07–M’08) received the B.E. degree in electronic engineering and the M.E. degree in signal and information processing from Xidian University, Xi’an, China, in 2001 and 2004, respectively, and the Ph.D. degree in elec-tronic engineering from the City University of Hong Kong, Hong Kong, in 2008.

He is currently a Senior Lecturer with the School of Electrical, Computer and Telecommu-nications Engineering, University of Wollongong, Wollongong, NSW, Australia, and also an Adjunct Associate Professor with the School of Electrical, Electronic and Computer Engineering, The University of Western Australia, Perth, WA, Australia. His research interests include signal processing and telecommunications. He was a recipient of the Australian Research Council’s Discovery Early Career Researcher Award.

DEFENG HUANG (S’02–M’05–SM’07) received the B.E.E.E. and M.E.E.E. degrees in electronic engineering from Tsinghua University, Beijing, China, in 1996 and 1999, respectively, and the Ph.D. degree in electrical and electronic engineer-ing from The Hong Kong University of Science and Technology, Hong Kong, in 2004.

He is currently a Professor with the School of Electrical, Electronic and Computer Engineering, The University of Western Australia. He serves as an Editor for the IEEE Wireless Communications Letters. He also served as an Editor for the IEEE Transactions on Wireless Communications from 2005 to 2011.

JINGWEI YIN was born in 1980. He is currently a Professor with the College of Underwater Acous-tic Engineering, Harbin Engineering University, China. His current research interests are underwa-ter acoustic engineering and underwaunderwa-ter acoustic communication.

MAOCHUN ZHENG was born in 1989. He is currently pursuing the Ph.D. degree with the Col-lege of Underwater Acoustic Engineering, Harbin Engineering University, China. He is currently a Visiting Scholar with The University of Western Australia. His current research interest is underwa-ter acoustic communication.